[Bug 697463] New: rtpsbcdebay: incorrect frame length calculation for dual channel.

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=697463
  GStreamer | gst-plugins-good | unspecified

           Summary: rtpsbcdebay: incorrect frame length calculation for
                    dual  channel.
    Classification: Platform
           Product: GStreamer
           Version: unspecified
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: Normal
         Component: gst-plugins-good
        AssignedTo: gstreamer-bugs at lists.freedesktop.org
        ReportedBy: andreas.fenkart at streamunlimited.com
         QAContact: gstreamer-bugs at lists.freedesktop.org
     GNOME version: ---


Created an attachment (id=240865)
 View: https://bugzilla.gnome.org/attachment.cgi?id=240865
 Review: https://bugzilla.gnome.org/review?bug=697463&attachment=240865

rtpsbcdebay: incorrect frame length calculation for dual  channel.

debug output in gst_rtp_sbc_depay_get_params reports this for IPad:
blocks:16 channels:2 subbands:8 bitpool:53 mode 0x02/dual channel

It seems to be wrong to multiply bitpool by number of channels.
I failed to find a proper spec clarifying that. The following RFC
has only a table about joint stereo, but not dual channel. The bitpool
values from IPad though, match those from stereo w/o joint, and these
are not multiplied by number of channels. checked by filling the numbers

http://tools.ietf.org/html/draft-ietf-payload-rtp-sbc-04#section-3

   +------------------------------------------------------------+
   | SBC encoder settings at Medium Quality                     |
   +--------------------------------+-------------+-------------+
   |                                |    Mono     | Joint Stereo|
   | Sampling frequency (kHz)       | 44.1 |  48  | 44.1 |  48  |
   | Bitpool value                  |  19  |  18  |  35  |  33  |
   | Resulting frame length (bytes) |  46  |  44  |  83  |  79  |
   | Resulting bit rate (kb/s)      | 127  | 132  | 229  | 237  |
   +--------------------------------+------+------+------+------+
   | SBC encoder settings at High Quality                       |
   +--------------------------------+-------------+-------------+
   |                                |    Mono     | Joint Stereo|
   | Sampling frequency (kHz)       | 44.1 |  48  | 44.1 |  48  |
   | Bitpool value                  |  31  |  29  |  53  |  51  |
   | Resulting frame length (bytes) |  70  |  66  | 119  | 115  |
   | Resulting bit rate (kb/s)      | 193  | 198  | 328  | 345  |
   +--------------------------------+------+------+------+------+
   + Other settings: Block length = 16, loudness, subbands = 8  |
   +------------------------------------------------------------+


Also pulseaudio is neither multiplying by number channels:
size_t sbc_get_frame_length(sbc_t *sbc)
{
[snip]
        subbands = sbc->subbands ? 8 : 4;
        blocks = 4 + (sbc->blocks * 4);
        channels = sbc->mode == SBC_MODE_MONO ? 1 : 2;
        joint = sbc->mode == SBC_MODE_JOINT_STEREO ? 1 : 0;
        bitpool = sbc->bitpool;

        ret = 4 + (4 * subbands * channels) / 8;
        /* This term is not always evenly divide so we round it up */
        if (channels == 1)
                ret += ((blocks * channels * bitpool) + 7) / 8;
                                  ^^^^^^^^ unnecessary, since it's 1
        else
                ret += (((joint ? subbands : 0) + blocks * bitpool) + 7) / 8;

        return ret;
}


... and audio plays perfect with the patch.

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.